A novel approach to measure awareness of memory decline and its ability to predict clinical progression in cognitively normal older adults

Abstract Background Heightened awareness and unawareness of memory decline are known to manifest at different stages in AD, and both have been shown to predict clinical progression. Awareness is often measured by the discrepancy between the informant’s and participant’s mean scores on a memory complaints questionnaire. A major limitation of this approach is that a participant may display heightened awareness on some items and unawareness on others, averaging out each’s effect. Here, we present a new item‐level methodological approach, allowing us to examine these two dimensions of awareness (heightened/unawareness). We aimed to compare all these measures and investigate their ability to predict clinical progression. Method This study included 438 ADNI participants with CDR global = 0 at the time of their baseline Everyday Cognition (ECog) measurement and at least two follow‐up CDR assessments (Tab.1). Disease progression was defined as the first instance of two consecutive CDR global≥0.5. A traditional awareness score was calculated by subtracting the mean average ECog score between the informant and participant complaints (Fig.1A). An unawareness score was generated by capping item‐level positive differences at zero and then averaging (Fig.1B). Its complementary heightened awareness score was generated by capping negative differences at zero (Fig.1C). For all three measures, baseline group differences were analyzed using Wilcoxon rank‐sum tests. Robust Cox regression models were run to test the association between each baseline measure and progression, adjusted for age, gender, and education. Result The progressing (N = 91) and stable (N = 347) groups demonstrated similar scores for traditional (p = .11) and heightened awareness (p = .72). A significant difference was found for unawareness, with progressors displaying more unawareness than stable participants (p<.001;Tab.1,Fig.1). Robust Cox regression models showed a significant association between greater risk of progression and unawareness (p<.001) and a marginal association with lower traditional scores (p = .06). There was not a significant association with heightened awareness (p = .75). Conclusion By separating its two dimensions, our novel approach allows us to refine the analysis of awareness and to determine each dimension’s respective contributions to future clinical progression. These findings provide evidence that unawareness rather than heightened awareness is associated with clinical progression and supports the utility of informant report of decline.